Excess estrogen exposure unopposed by progesterone is associated with increased risk of endometrial cancer. Polymorphisms in genes involved in estrogen metabolism influence the levels of these hormones and may be associated with an altered risk of endometrial cancer; however few studies have been conducted to date. Endometrial cancer is particularly worth studying because it is the most estrogen associated tumor and thus modest effects of hormone-metabolizing genes may be more easily detectable. To this end, we propose to continue to study polymorphisms in selected candidate genes and their relation to endometrial cancer using a large well-characterized cohort, the Nurses' Health Study (NHS), and to add a second large cohort, the Women's Health Study (WHS). The addition of this second cohort will substantially increase our power to assess whether polymorphisms in genes that metabolize estradiol and their receptors are predictive of future endometrial cancer risk. We will assess the functional significance of the variant alleles in the NHS by correlating these variants with plasma hormone levels. We will assess whether functional polymorphisms and haplotypes in StAR, CYP11A1, 17aHSD type 4, and 3a-HSD type 1 and 2. SHBG, ESR1 and AR are associated with endometrial cancer risk. We will seek to replicate associations we observed between endometrial cancer risk, and CYP17 and CYP19. Finally, we will quantify the association with endometrial cancer and test whether these associations are modified by established endometrial cancer risk factors such as body mass index and exogenous hormone use. Our study will be among the few studies able to prospectively examine these issues in two large defined cohorts with complete ascertainment of incident cases and comprehensive prospective information on other endometrial cancer risk factors. With the addition of the WHS we will have a total of 955 invasive endometrial cancer cases. We will have >85% power to detect a relative risk 1.50 or greater for the main effects of most of the genotypes of interest.